Not all Infrastructure-as-a-Service (IaaS) clouds are created equal.

Some solutions allow administrators to build images from a template. Others require that images (virtual machines) be constructed outside the product and then imported.
Some solutions can only create images for and deploy them to platforms and clouds running specific proprietary technology stacks. Others offer choice and portability.

However, for the purpose of this discussion let’s leave aside differences in how images are defined, built and deployed. Instead, let's consider ongoing management.

As the figure shows, without a management component that completes the circle from the running application back to the original content, we believe that an IaaS is essentially taking a fire-and-forget approach after images are launched into a cloud or clouds.
A number of IaaS solutions do manage the container, i.e. the image or VM, to a greater or lesser degree. The thinking is that if some problem with a running image is detected, the image can be shutdown and a new one deployed. This way of thinking reflects how certain types of applications, typically newer ones, are "stateless"--which is to say that they don't store data unique to the running instance—and therefore they can just be restarted with relatively little disruption.

While this approach makes sense in some cases, it has its drawbacks. The first is that many applications are not, in fact, stateless. The second is that, without the ability to monitor the content within the container in real-time, it can be difficult to detect whether some sort of "drift" has, in fact, happened.

By contrast, Red Hat CloudForms manages the content as well as the container. It is designed to operationally manage running systems across physical, virtual and cloud environments. By doing so, it offers more continuous compliance of content and configurations (as well as entitlements for Red Hat products) consistent with the definitions used to build the image.
The continuous management of content handles two primary use cases.

In the first use case, the system administrator updates the definition of an application and wants to push that update out to a running production system. While the admin could simply reprovision and relaunch the resource, continuous content management adds the option of doing a live update that does not require rebuilding and restarting the running instance.

In the second case, a developer might modify an instance by updating a library, bringing that instance out of compliance with the application definition. Such a change would be detected and the administrator alerted. The admin would then decide whether to remediate by updating the library to the supported version.
This important advantage that CloudForms enjoys over many more rudimentary cloud management offerings lies in its ability to do content, configuration and compliance lifecycle management. This automated control over applications while they are running means that administrators can have greater confidence that their infrastructure is under control at all times and not just at the instant a new image starts up. And even if all you’re managing today is physical servers, Red Hat gives you a highly effective tool to manage those systems while preparing you to transition your applications to the cloud when you are ready. Managing applications within an operating system "container" is a tough challenge even in the absence of cloud computing, yet it’s ignored by many IaaS solutions.

Runtime management at the application level means that "drifts" in the configuration of applications can be detected and more quickly remediated without restarting the application—a key requirement for many types of critical business applications.

About the author